03-27-2018 07:37 AM - edited 03-05-2019 10:10 AM
We recently went through a few upgrades at our CORP center, make a long story short we are seeing high bandwidth utilization and latency at remote branch locations when a user opens up their Outlook client to download their mailbox(caching mode is off by default).
The exchange backend sits at our CORP center which connects to our MPLS (50Mbps), the branch however only have a 6Mbps pipe to the MPLS. I've played around with applying a policy-map at the branch but I still see high utilization. Would one be better off applying a policy-map close the Exchange backend before it gets on to the MPLS or does anyone have any suggestions?
03-27-2018 08:04 AM
03-27-2018 08:38 AM
We only have currently a policy-map across all locations is 40% for VOIP traffic. Our WAN MPLS provider is using standard QoS markings.
So if I understand you correctly(we wanted to use 20% of the line for Exchange traffic) I would implement a policy-map at a remote location with a bandwidth of 6mbps using the following:
ip access-list extended EXCHANGE
permit ip any host Exchange_Servers
class-map EXCHANGE
match access-group name EXCHANGE
policy-map QoS_OUT
class EXCHANGE
police 1200000 conform-action transmit exceed-action drop
On the CORP location use something similar with 20% of the 50mbps line used for Exchange traffic?
03-27-2018 09:05 AM - edited 03-28-2018 07:16 AM
" Our WAN MPLS provider is using standard QoS markings."
Oh, what standard, the RFC? What exactly do they do per marking.
What are your devices, and their IOS versions, being used both at CORP and your branches. Could you post what you're using now, for policies, on both sides?
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide